Strategy Switches Research Articles

Abrupt and spontaneous strategy switches emerge in simple regularised neural networks.

Humans sometimes have an insight that leads to a sudden and drastic performance improvement on the task they are working on. Sudden strategy adaptations are often linked to insights, considered to be a unique aspect of human cognition tied to complex processes such as creativity or meta-cognitive reasoning. Here, we take a learning perspective and ask whether insight-like behaviour can occur in simple artificial neural networks, even when the models only learn to form input-output associations through gradual gradient descent. We compared learning dynamics in humans and regularised neural networks in a perceptual decision task that included a hidden regularity to solve the task more efficiently. Our results show that only some humans discover this regularity, and that behaviour is marked by a sudden and abrupt strategy switch that reflects an aha-moment. Notably, we find that simple neural networks with a gradual learning rule and a constant learning rate closely mimicked behavioural characteristics of human insight-like switches, exhibiting delay of insight, suddenness and selective occurrence in only some networks. Analyses of network architectures and learning dynamics revealed that insight-like behaviour crucially depended on a regularised gating mechanism and noise added to gradient updates, which allowed the networks to accumulate "silent knowledge" that is initially suppressed by regularised gating. This suggests that insight-like behaviour can arise from gradual learning in simple neural networks, where it reflects the combined influences of noise, gating and regularisation. These results have potential implications for more complex systems, such as the brain, and guide the way for future insight research.

Equilibrium seeking of higher-order networks under facet cover constraints.

Agent intelligence involves specific requirements for social attributes. Intelligent agents make their decisions based on the groups they are part of, tend to satisfy co-members, and enlarge their own benefits. A fundamental question is whether this form of subgroup decision-making accommodate each individual's preferences. In this paper, we examine the evolution of an anticoordination game on a higher-order network in the form of a simplicial complex in relation to the facet cover problem, which ensures that each subgroup yields a positive benefit. We introduce and apply the facet update rule to regulate nodes' group-based interactions. We identify the payoff parameter condition that a strict Nash equilibrium (SNE) satisfies a facet cover. The proposed facet update rule enables the activated facet to reach a facet equilibrium, and all nodes would converge to an SNE with no more than 2m strategy switches, where m is the number of nodes in the simplicial complex. Additionally, we analyze the convergence of the asynchronous update rule, which can be seen as a special case of the facet update rule. Our simulations and extensive examples reveal that the SNE achieved by the facet update rule, on average, covers fewer nodes compared to the asynchronous update rule.

Asynchronous best-response dynamics of networked anti-coordination game with payoff incentives

How to understand and model the spatial collective decision-making behaviors in the evolutionary game context is an open problem, which attracts burgeoning interdisciplinary studies in the fields of graph theory and evolutionary game theory. This is a newly emerging research topic with wide applications under many real engineering, social and natural nonlinear systems. In this paper, we investigate the asynchronous best-response dynamics of anti-coordinating agents by developing two types of payoff incentive mechanisms, i.e., reward and punishment. We start from characterizing the equilibria of the best-response dynamics based on various types of graph structures. By offering relevant payoff incentives to targeted nodes, we further explore the global outcomes of strategy evolution. Current results suggest that both the reward and punishment can cause the strategy switches of targeted nodes, thus to induce a cascading effect of the evolutionary dynamics in the spatial gaming systems. Given the unlimited and limited budget cases for the payoff incentives, we calculate the optimal solutions for the added incentives, which facilitates the prediction on how the system dynamics be influenced by the proposed incentive mechanisms.

Research on Improved Droop Control of AC-DC Microgrid Bidirectional Converter Based on Power

The bidirectional converter in AC / DC hybrid microgrid is an important link to balance the power on both sides of AC and DC, Its control method is extremely important. This paper proposes an improved droop control strategy for bidirectional converter using DC-side power amount control. This control can autonomously and smoothly switch in rectifyer, inverter and idle modes. Utilizing DC-side power control avoids normalized calculations. Strategy meets two modes of island and grid connection, eliminating the need for multiple control strategy switches. Simulation results show that the DC bus voltage and AC frequency are near the rated value, which guarantees the stable operation of the AC-DC hybrid microgrid.

Migrant pool model of dispersal explains strong connectivity of Avicennia officinalis within Sundarban mangrove areas: Effect of fragmentation and replantation

The ability of mangrove trees to reproduce and disperse in highly dynamic tidal environments is reflected in the contemporary genetic diversity and structure of their populations. Resilience and establishment may occur through either local recruitment (propagule pool model) or over longer distances with mixed origins (migrant pool model), thereby affecting the connectivity between natural populations. Anthropogenic activities such as encroachment and fragmentation or from mitigating reforestation could cause strategy switches and thus the connectivity and genetic structure of populations. In this study, we considered Avicennia officinalis, a dominant component of the Sundarban mangrove forest in Bangladesh to estimate dispersal dynamics over various distances within estuaries. A total of 423 individuals of A. officinalis, distributed over thirteen sites in the Sundarbans were sampled at both large and fine-scaled spatial level, the latter for comparison of natural and dense with fragmented sites. Up to 62 alleles showed a comparable level of allele diversity and gene diversities at both juvenile and adult life stages, assuming overall high levels of outcrossing. Significant inbreeding levels and high kinship values were solely observed at a reforested A. officinalis site. Genetic variables revealed high levels of connectivity between sites at close vicinity, with only a weak structure over larger geographic distances, between natural core Sundarban and coastal edge populations. Kinship coefficients were mostly low though significant up to several kilometers along estuarine branches, thereby suggesting propagule dispersal beyond each site. A fine-scaled analysis of kinship levels estimated for fragmented and natural forest plots along both sides of a river revealed that fragmentation induced higher kinship levels over 25–50 m distance, whereas dense natural populations were composed of less related individuals at shortest distances. A. officinalis populations might stay resilient through a mixture of unrelated propagules, clearly adhering a migrant pool model in natural sites and propagule pool model in fragmented sites. A poor gene diversity of planted trees might jeopardize the sustainability of mangrove plantation practices. Furthermore, high levels of diversity and connectivity indicated that the core Sundarban populations are a relevant source of germplasm in afforestation programmes.

Behavioral Strategy Determines Frontal or Posterior Location of Short-Term Memory in Neocortex

The location of short-term memory in mammalian neocortex remains elusive. Here we show that distinct neocortical areas maintain short-term memory depending on behavioral strategy. Using wide-field and single-cell calcium imaging, we measured layer 2/3 neuronal activity in mice performing a whisker-based texture discrimination task with delayed response. Mice either deployed an active strategy-engaging their body toward the approaching texture-or passively awaited the touch. Independent of strategy, whisker-related posterior areas encoded choice early after touch. During the delay, in contrast, persistent cortical activity was located medio-frontally in active trials but in a lateral posterior area in passive trials. Perturbing these areas impaired performance for the associated strategy and also provoked strategy switches. Frontally maintained information related to future action, whereas activity in the posterior cortex reflected past stimulus identity. Thus, depending on behavioral strategy, cortical activity is routed differentially to hold information either frontally or posteriorly before converging to similar action.

Trial-by-trial identification of categorization strategy using iterative decision-bound modeling.

Identifying the strategy that participants use in laboratory experiments is crucial in interpreting the results of behavioral experiments. This article introduces a new modeling procedure called iterative decision-bound modeling (iDBM), which iteratively fits decision-bound models to the trial-by-trial responses generated from single participants in perceptual categorization experiments. The goals of iDBM are to identify: (1) all response strategies used by a participant, (2) changes in response strategy, and (3) the trial number at which each change occurs. The new method is validated by testing its ability to identify the response strategies used in noisy simulated data. The benchmark simulation results show that iDBM is able to detect and identify strategy switches during an experiment and accurately estimate the trial number at which the strategy change occurs in low to moderate noise conditions. The new method is then used to reanalyze data from Ell and Ashby (2006). Applying iDBM revealed that increasing category overlap in an information-integration category learning task increased the proportion of participants who abandoned explicit rules, and reduced the number of training trials needed to abandon rules in favor of a procedural strategy. Finally, we discuss new research questions made possible through iDBM.

Setbacks in the process of assimilation of problematic experiences in two cases of emotion-focused therapy for depression

Objective: Research on the assimilation model has suggested that psychological change takes place in a sequence of stages punctuated by setbacks, that is, by transient reversals in the developmental course. This study analyzed such setbacks in one good outcome case and one poor outcome case of Emotion-focused therapy (EFT) for depression. Method: Intensive analyses of five transcribed sessions from each case identified 26 setbacks in the good outcome case and 27 in the poor outcome case. The reason for each setback was classified into one of four categories: balance strategy, exceeding the therapeutic zone of proximal development either induced by the therapist (ZPD-T) or induced by the client (ZPD-C), or spontaneous switches. Results: In the good outcome case the most frequent reasons for setbacks were balance strategy and spontaneous switches, whereas in the poor outcome case the most frequent reason for setbacks was ZPD-T. Conclusions: As in previously studied therapies, setbacks in EFT, usually represent productive work on relatively less advanced strands of the client's major problems. Results point to the importance of the therapist attending to the limits of the client's therapeutic ZPD.

Relationships Between Strategy Switching and Strategy Switch Costs in Young and Older Adults: A Study in Arithmetic Problem Solving

Background/Study Context: This study investigated age-related differences in within-item strategy switching (i.e., revising initial strategy choices to select a better strategy while solving a given problem) and in strategy switch costs (i.e., longer latencies when participants switch strategies than when they do not switch strategy during strategy execution).Methods: In a computational estimation task, participants had to give approximate products to two-digit multiplication problems (e.g., 41 × 67) while rounding up (i.e., do 50 × 70 for 41 × 67) or rounding down (i.e., do 40 × 60 for 41 × 67) operands to their nearest decades. After executing a cued strategy during 1000 ms, participants had the possibility to switch to another strategy (or repeat the same strategy) in a selection condition. In an execution condition, participants were forced to repeat the same strategy or to switch to another strategy.Results: It was found that (1) older adults were less able than young adults to switch strategy after starting to execute a cued strategy (36.1% vs. 45.8%); (2) older adults showed larger switch costs than young adults (422 vs. 223 ms); and (3) strategy switches and strategy switch costs correlated in older adults but not in young adults.Conclusion: These findings have important implications for our understanding of the mechanisms underlying within-item strategy switching and aging effects on these mechanisms as well as, more generally, of strategic variations during cognitive aging.

Dynamic complementarities in innovation strategies

Using a panel of Irish manufacturing plants over the period 1991–2008 we test for dynamic complementarities in the joint use of internal R&D and external knowledge sources. We find little evidence, either from considering successive cross-sectional waves of comparable surveys, or in terms of the strategy switch choices of specific plants, that there has been a systematic move towards the joint use of internal and external knowledge in innovation. We then test formally for the presence of complementarities in the joint use of internal R&D and external innovation linkages. In static terms we find no evidence of complementarity, but in dynamic terms find evidence that strategy switches by individual plants towards a more ‘open’ strategy are accompanied by increased innovation outputs.

Thalamic and parietal brain morphology predicts auditory category learning

Auditory categorization is a vital skill involving the attribution of meaning to acoustic events, engaging domain-specific (i.e., auditory) as well as domain-general (e.g., executive) brain networks. A listener's ability to categorize novel acoustic stimuli should therefore depend on both, with the domain-general network being particularly relevant for adaptively changing listening strategies and directing attention to relevant acoustic cues. Here we assessed adaptive listening behavior, using complex acoustic stimuli with an initially salient (but later degraded) spectral cue and a secondary, duration cue that remained nondegraded. We employed voxel-based morphometry (VBM) to identify cortical and subcortical brain structures whose individual neuroanatomy predicted task performance and the ability to optimally switch to making use of temporal cues after spectral degradation. Behavioral listening strategies were assessed by logistic regression and revealed mainly strategy switches in the expected direction, with considerable individual differences. Gray-matter probability in the left inferior parietal lobule (BA 40) and left precentral gyrus was predictive of “optimal” strategy switch, while gray-matter probability in thalamic areas, comprising the medial geniculate body, co-varied with overall performance. Taken together, our findings suggest that successful auditory categorization relies on domain-specific neural circuits in the ascending auditory pathway, while adaptive listening behavior depends more on brain structure in parietal cortex, enabling the (re)direction of attention to salient stimulus properties.

Understanding human high-level spatial memory: An ACT-R model to integrate multi-level spatial cues and strategies

Abstract The ability to process and use spatial knowledge is a basic cognitive ability. Two human navigation strategy types (map-based and route-based) relying on two different knowledge representations have been frequently observed. These studies suggest that the first strategy uses a sequential representation and the second uses a hierarchical cluster-based representation. These studies also suggest that humans also routinely use hybrid strategies, and that the ratio between cognitive load and relative utility mediated by situational factors influences, and when modeled, could successfully predict strategy choice. We created an ACT-R model to test these hypotheses by simulating navigation strategies, strategy choices, and strategy switches. This model deepens the empirical findings by defining more clearly the memory mechanisms involved in generating the basic representation types, and by positing a theory of interaction between these types based on ACT-R’s associative declarative memory. We believe that such a work provides a concrete example on principles of these biological theories can be implemented and used in cognitive architectures.

The differential involvement of the prelimbic and infralimbic cortices in response conflict affects behavioral flexibility in rats trained in a new automated strategy-switching task

To assess the role of the prelimbic (PL) and infralimbic (IL) cortices in mediating strategy switching, rats were trained in a new automated task in a Y-maze allowing a careful analysis of rats' behavior. In this situation, rats can only use two egocentric (Right, Left) and two visual (Light, Dark) strategies. In the first experiment, rats with PL, IL, or PL/IL lesions were compared with sham-operated rats when trained to reach a criterion of 10 consecutive correct responses with a light strategy before being trained with a response strategy (rule shifting), and finally with the reversed response strategy (reversal). In the second experiment, sham-operated and PL-lesioned rats had their first two strategy switches in the reverse order, which was followed by a second rule shifting and reversal. The results indicate that lesions did not affect initial acquisition, but impaired the first rule shifting and reversal. Thorough analyses of rats' performance indicate that lesioned rats were still able to demonstrate some behavioral flexibility but have difficulties in solving response conflicts, which in turn may affect behavioral flexibility. Both areas were differentially involved in the resolution of response conflict, with the IL involved in the choice of strategy previously known to be nonvalid, and the PL in the selection and maintenance of that strategy.

Lateralized cerebellar contributions to word generation: a phonemic and semantic fluency study.

Impairment on verbal fluency tasks has been one of the more consistently reported neuropsychological findings after cerebellar lesions, but it has not been uniformly observed and the possible underlying cognitive basis has not been investigated. We tested twenty-two patients with chronic, unilateral cerebellar lesions (12 Left, 10 Right) and thirty controls on phonemic and semantic fluency tasks. We measured total words produced, words produced in the initial 15 seconds, errors and strategy switches. In the phonemic fluency task, the right cerebellar lesion (RC) group produced significantly fewer words compared to the left cerebellar lesion (LC) group and healthy controls, particularly over the first 15 seconds of the task with no increase in errors and significantly fewer switches over the entire task. In the semantic fluency task there was only a modest decrease in total words in the RC group compared to controls. RC lesions impair fluency with many of the same performance characteristics as left prefrontal lesions. This supports the hypotheses of a prefrontal-lateral cerebellar system for modulation of attention/executive or strategy demanding tasks.

Lateralized Cerebellar Contributions to Word Generation: A Phonemic and Semantic Fluency Study

Impairment on verbal fluency tasks has been one of the more consistently reported neuropsychological findings after cerebellar lesions, but it has not been uniformly observed and the possible underlying cognitive basis has not been investigated. We tested twenty-two patients with chronic, unilateral cerebellar lesions (12 Left, 10 Right) and thirty controls on phonemic and semantic fluency tasks. We measured total words produced, words produced in the initial 15 seconds, errors and strategy switches. In the phonemic fluency task, the right cerebellar lesion (RC) group produced significantly fewer words compared to the left cerebellar lesion (LC) group and healthy controls, particularly over the first 15 seconds of the task with no increase in errors and significantly fewer switches over the entire task. In the semantic fluency task there was only a modest decrease in total words in the RC group compared to controls. RC lesions impair fluency with many of the same performance characteristics as left prefrontal lesions. This supports the hypotheses of a prefrontal-lateral cerebellar system for modulation of attention/executive or strategy demanding tasks.

Double dissociation and hierarchical organization of strategy switches and reversals in the rat PFC.

Prelimbic-infralimbic cortex (PL-IL) and orbitofrontal cortex (OFC) influence behavioral flexibility in the rat. The authors tested the effects of PL-IL or OFC infusion with the GABA agonist muscimol in the context of 2 flexible responding tasks: strategy switching and reversal. Muscimol infusion into PL-IL impaired retention of strategy switches but not reversals, whereas muscimol infusion into OFC impaired retention of reversals but not switches. However, whereas training in repeated reversals did not remove the requirement of PL-IL for switch retention (E. L. Rich & M. L. Shapiro, 2007), training in repeated switches did remove the requirement of OFC for reversal retention. Thus, activity during strategy switches was sufficient to initiate learning and remove the requirement of OFC in later reversals.

Models of probabilistic category learning in Parkinson’s disease: Strategy use and the effects of L-dopa

Probabilistic category learning (PCL) has become an increasingly popular paradigm to study the brain bases of learning and memory. It has been argued that PCL relies on procedural habit learning, which is impaired in Parkinson’s disease (PD). However, as PD patients were typically tested under medication, it is possible that levodopa (L-dopa) caused impaired performance in PCL. We present formal models of rule-based strategy switching in PCL to re-analyse the data from [Jahanshahi, M., Wilkinson, L., Gahir, H., Dharminda, A., & Lagnado, D.A., (2009). Medication impairs probabilistic classification learning in Parkinson’s disease. Manuscript submitted for publication] comparing PD patients on and off medication (within subjects) to matched controls. Our analysis shows that PD patients followed a similar strategy switch process as controls when off medication, but not when on medication. On medication, PD patients mainly followed a random guessing strategy, with only few switching to the better Single Cue strategies. PD patients on medication and controls made more use of the optimal Multi-Cue strategy. In addition, while controls and PD patients off medication only switched to strategies which did not decrease performance, strategy switches of PD patients on medication were not always directed as such. Finally, results indicated that PD patients on medication responded according to a probability matching strategy indicative of associative learning, while the behaviour of PD patients off medication and controls was consistent with a rule-based hypothesis testing procedure.

Parallel acquisition of allocentric and egocentric navigation strategies in the human brain

Event Abstract Back to Event Parallel acquisition of allocentric and egocentric navigation strategies in the human brain Kinga Igloi1*, Christian Doeller2, Alain Berthoz1, Neil Burgess2 and Laure Rondi-Reig3 1 LPPA College de France, France 2 ICN UCL, France 3 CNRS, NPA UPMC, France Navigation in a complex environment can rely on the use of different spatial strategies. We focused on allocentric (or encoding interrelationships among environmental cues, movements, and the location of the goal) and sequential egocentric (or sequences of body turns associated with specific choice points) strategies. Traditional questions on learning speed and hierarchy between strategies are still debated. Here we adapted a novel spatial paradigm, developed for rodents, first in a behavioral study then for use with fMRI in humans and asked the question if allocentric and egocentric strategies are acquired in parallel. We first investigated spontaneous or imposed choice and switches between these two spatial strategies in a behavioral study: our results show bi-directional strategy switches, suggesting no temporal or performance-related hierarchy between strategies. Second, using fMRI, we showed that allocentric navigation activates a neural network comprising the right hippocampus, ventral and dorsal striatum and the precuneus; whereas sequential egocentric navigation involves the left hippocampus, ventral and dorsal striatum and the posterior insula. We further report parallel activation of the main neural structures identified in both strategies (bilateral hippocampus, precuneus, striatum, insula) during the initial learning process, thus confirming parallel acquisition of these two strategies. Additionally, medial prefrontal activation could be linked to cognitive control and monitoring of available strategies during learning. Interestingly, both strategies can be identified in hippocampal activity prior to their expression in behavior. We show that, activation of the right hippocampus supports the allocentric strategy (memory for places), while activation of the left hippocampus supports the sequential egocentric strategy (memory for temporal sequences), which can be used to support both spatial navigation and episodic memory more generally. Conference: 41st European Brain and Behaviour Society Meeting, Rhodes Island, Greece, 13 Sep - 18 Sep, 2009. Presentation Type: Poster Presentation Topic: Poster presentations Citation: Igloi K, Doeller C, Berthoz A, Burgess N and Rondi-Reig L (2009). Parallel acquisition of allocentric and egocentric navigation strategies in the human brain. Conference Abstract: 41st European Brain and Behaviour Society Meeting. doi: 10.3389/conf.neuro.08.2009.09.177 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 10 Jun 2009; Published Online: 10 Jun 2009. * Correspondence: Kinga Igloi, LPPA College de France, Paris, France, kinga.igloi@college-de-france.fr Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Kinga Igloi Christian Doeller Alain Berthoz Neil Burgess Laure Rondi-Reig Google Kinga Igloi Christian Doeller Alain Berthoz Neil Burgess Laure Rondi-Reig Google Scholar Kinga Igloi Christian Doeller Alain Berthoz Neil Burgess Laure Rondi-Reig PubMed Kinga Igloi Christian Doeller Alain Berthoz Neil Burgess Laure Rondi-Reig Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

Binary Games with State Dependent Stochastic Choice

This paper studies a stochastic equilibrium selection model for binary coordination games. Players switch strategies stochastically so that the mistake probabilities are fully dependent on the population states. A probabilistic behavior is said to be aspiration (imitation, resp.) oriented if strategy switches are mainly driven by the aspiration (imitation, resp.) effect. In general, a strategy switch by one player generates externalities on others. Strategies in a coordination game can be classified according to the relative magnitude of their externality effects. It is shown that the selection outcome for a linear coordination game is determined in a specific way by the balance of the risk dominance, the aspiration/imitation, and the externality effects. It is also shown that an aspiration (imitation, resp.) oriented behavior tends to select payoff dominant (maxmin, resp.) equilibrium and that risk dominant equilibrium is always selected if and only if the aspiration and the imitation effects exactly cancel each other out, which in turn makes the selection process insensitive to externality effects. Journal of Economic Literature Classification Number: C72.